The present disclosure describes techniques for generating a mask frame data segment corresponding to a video frame. The disclosed techniques include obtaining a frame of a video; identifying a main area of the frame using an image segmentation algorithm; and generating a mask frame data segment corresponding to the frame based on the main area of the frame, wherein the generating a mask frame data segment corresponding to the frame based on the main area of the frame further comprises generating the mask frame data segment based on a timestamp of the frame in the video, a width and a height of the main area of the frame.
Legal claims defining the scope of protection. Each claim is shown in both the original legal language and a plain English translation.
2. The method of claim 1, wherein the identifying a main area of the frame using an image segmentation algorithm further comprises identifying the main area of the frame by a semantic segmentation model, and the semantic segmentation model comprises at least two feature extraction models, at least one feature enhancement layer, and a classification layer.
4. The method of claim 1, wherein the identifying a main area of the frame using an image segmentation algorithm further comprises identifying the main area of the frame by an instance segmentation model.
The invention relates to image processing, specifically to methods for identifying and analyzing regions of interest within a frame, such as in video or image analysis applications. The problem addressed is the need for accurate and efficient segmentation of a main area within a frame, which is crucial for tasks like object detection, scene understanding, or automated content analysis. The method involves using an image segmentation algorithm to isolate the main area of the frame. Specifically, an instance segmentation model is employed to distinguish the main area from other regions. Instance segmentation not only classifies pixels but also differentiates between individual objects or regions within the frame, providing precise boundaries for the main area. This approach improves accuracy over traditional segmentation techniques by leveraging deep learning models that can handle complex scenes with multiple overlapping objects. The instance segmentation model is trained to recognize and segment the main area based on features such as shape, texture, or contextual information. The output is a segmented frame where the main area is clearly delineated, enabling further processing like object tracking, anomaly detection, or content-based retrieval. This method is particularly useful in applications requiring high precision, such as medical imaging, autonomous driving, or surveillance systems. The use of instance segmentation ensures that the main area is accurately identified even in cluttered or dynamic environments.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
May 13, 2020
October 18, 2022
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.